Feeding digital human models: a motion capture-based flexible methodology for simulating operator actions in high-risk environments

Operators working in critical industrial environments, such as oil & gas facilities, chemical plants, or confined spaces, are frequently exposed to hazardous conditions, especially during emergency interventions or failures caused by technical malfunctions and cyberattacks. In such scenarios, human intervention is often necessary to detect the anomaly and perform corrective actions. However, direct operator involvement may be extremely dangerous due to environmental risks, access limitations, and strict safety regulations. Traditional training methods require physical presence in high-risk areas and can be impractical or unsafe. This paper proposes a modular methodology to feed Digital Human Models (DHMs) with real motion capture (MoCap) data, enabling the simulation of operator behavior in high-risk environments. The approach involves recording, in a safe laboratory setting, basic actions (BAs) the operator can perform while working on site, using MoCap technology, which are then processed and concatenated to build fully digital operative scenarios. The modular methodology allows the creation of a flexible and scalable database of BAs from which digital operative scenarios can be built, supporting training, safety validation, and human–machine interaction (HMI) analysis without exposing humans to risks. The methodology is tested in a lab prototype of an oil & gas plant, replicating an operational anomaly and comparing the digitally simulated scenario to a physically recorded one. Results show comparable behavior patterns and timing, proving the system’s ability to reproduce real-world operations. This approach offers a promising alternative to traditional training and assessment methods, contributing to safer, effective, and proactive risk management in high-risk industrial contexts.